Sertain Problematic Aspects of Fuzzy PID Regulation

Abstract

In recent years in view of the fact that the interest in the technologies of artificial intelligence in control systems by technical objects and technological processes are aroused, that fuzzy controllers find use more often. The given regulators carry out the process of the development of controlling actions on the basis of fuzzy logic, the application of which provides the design of control systems capable of functioning in conditions of incompleteness and fuzziness of knowledge about the dynamics of the control object. Fuzzy regulators have opened a new direction in the area of automatic control and according to many experts have a promising future. The given paper analyzes the dynamic characteristics and algorithmic singularities of fuzzy PID-regulation systems, compares the quality of clear and fuzzy regulation. The conducted analysis allows to state the absence of any real advantages of fuzzy regulators in comparison with classic clear regulators. Moreover, it is possible to distinguish a number of problematic aspects of fuzzy regulation methodology, important for practical automation:fuzzy control algorithms are much more complex than traditional clear regulation algorithms;the thesis about the advantages of fuzzy regulators seems to be reasonable, since each such regulator can be replaced by a more efficient and structurally less complex clear regulator;the thesis that on the basis of fuzzy approach it is possible to synthesize working systems of regulation without a priori knowledge and pre-project inspection of dynamic properties of objects of regulation is disputable.fuzzy approach is purely empirical and does not allow to solve the problems of stability, dynamic quality and robustness of the synthesized regulation systems at the theoretical level;fuzzy control algorithms are not applicable to complex dynamic objects. In particular, it concerns multi-connected objects of regulation and objects with delay;the methodology of fuzzy regulation does not allow to solve the important issues of optimization of regulation processes for engineering practice.